Tensile Bolt Fatigue Question

[TEDstruc]

Application: Fully Pre-tensioned Bolts subject to tensile cyclical loads. I am looking specifically at AISC 360-10 today since that is the edition I have at the office (I have 360-15 at home), although I think 360-15 is similar.

Appendix 3, section 3.4, states "For joints in which the material within the grip is limited to steel and which are tensioned to the requirements of Table J3.1 (Fully Pretensioned), an analysis of the relative stiffness of the connected parts and bolt shall be permitted to be used to determine the tensile stress range in the pretensioned bolts due to the total service live load...etc...."

However Table A-3.1, section 8.5 (the only section applicable to bolts in tension), applies only to non-pretensioned bolts.

Commentary on appendix 3, section C3.4 goes on to say that "Fatigue resistance of bolts subject to tension is predictable in the absence of pretension and prying action....etc....To limit the uncertainties regarding prying action on the fatigue of pretensioned bolts in details which introduce prying, the design stress range provided in Table A-3.1 is appropriate for extended cyclic loading only if the prying induced by the applied load is small." Okay.... I guess we should just size our connecting elements to eliminate all prying in this application...since there seems to be no indication on what "small" prying is...

RCSC Section 4.2 lists "Joints with ASTM A325 or F1852 bolts that are subject to tensile fatigue" under joint applications required to be fully pretensioned.

So to summarize, we are required to fully pretension the bolts, but the allowable stress range is only applicable for non-pretensioned fasteners, or fully pretensioned fasteners where the joints see only "small" prying.

Furthermore, What is the allowable total bolt tensile stress when bolts are fully pre-tensioned? For non-pretensioned, allowable stress is 45 ksi for A325 bolts, but a fully pretensioned bolt is pretensioned to about 64 ksi... so what is the allowable total tensile stress for a PT bolt? I understand that pretensioning reduces cyclic stress range if no prying is introduced, but the externally applied cyclic load is still stressing the bolt to some total stress value above the initial pretension stress....

 

[ProgrammingPE]

AISC 360-16 updated Table A-3.1, Section 8.5 to include pretensioned bolts.

RCSC 2014, Section 5.5, which discusses tensile fatigue, says "The connected parts shall be proportioned so that the calculated prying force does not exceed 30 percent of the externally applied load." This can be used as the limit of what is a "small" prying force.

When you calculate the stress range for a pretensioned bolt, you can use 20% of the actual load per the end of AISC 360-16, Section 3.4. Alternatively, you can determine the actual effect of the pretensioning on the bolts using the method in Section 6.2 of Design Guide 17 where you compare the area of the bolts and plate to determine how much of the external load goes to the bolts and how much just decreases the compression between the plates.

RCSC 2014, Table 5.2 lists a maximum tensile stress of 45 ksi, but the footnote specifically excludes the pretensioning.

I hope these references help.

Structural Engineering Software:

http://www.structuralcentral.com

Structural Engineering Videos:

http://www.youtube.com/channel/UCOj2mXXf3ZbZtgxTc6BtkGg

 

[TEDstruc]

Thanks ProgrammingPE, that's helpful.

So for infinite fatigue life... the threshold stress range specified in appendix 3 (7 ksi) is the allowable stress range for Cyclic Stress, while the stress listed in RCSC Table 5.2 (31 ksi for 500,000 cycles or more) is the max stress allowed from total load (DL + LL + Cyclic), correct? This all assumes that prying force is 30% or less of total tension on the bolt.

 

[rb1957]

so it looks to me as someone not in your business that you discount the material strength with static safety factors, so your allowable strength is very low (45ksi) but then these factors aren't appropriate for fatigue design so you preload to 1/2 yield (or some such) something like 60+ksi ?

another day in paradise, or is paradise one day closer ?

 

[ProgrammingPE]

TEDstruct, that sounds correct. The Appendix 3 and RCSC methods are different (stress range vs max stress) which can create some confusion. Make sure you still include the prying force (if it exists) even if it is less than 30% of the total force. Also, Appendix 3 uses the net area while RCSC uses the nominal area when calculating the stresses.

rb1957, the bolts get pretensioned to at least 70% of the yield stress and there may even be some yielding when this occurs, but it is a controlled process so rupture or excessive yielding will not occur. When the bolts get designed for the applied loads, we limit the tensile stress to about 50% of the yield stress due to uncertainty in the actual loads and material strengths.

Structural Engineering Software:

http://www.structuralcentral.com

Structural Engineering Videos:

http://www.youtube.com/channel/UCOj2mXXf3ZbZtgxTc6BtkGg

 

[rb1957]

I was trying to make sense of preload stresses exceeding allowables. I understand about 70% yield preload, noting that most preloading has a scatter of +-33%.

"70% preload" is IMHO a "lazy" one size fits all approach ... but I acknowledge the need for (and personal use of) "lazy" approaches (ie, I'm calling myself "lazy" too).

another day in paradise, or is paradise one day closer ?

 

[TEDstruc]

Thanks again ProgrammingPE, that is how I have always done these checks based on examples from a more experienced engineer, but I never could quite pin down all of the reasons and code references. That's what had me going thru all of the pertaining sections of the code yesterday and getting frazzled.